Overview
AI Commons are data platforms designed to support small to midscale AI models such as those created by the M3 initiative. They are intended to make powerful AI tools available alongside core data commons capabilities. Core services for an AI commons include:
- FAIR data
- Embedding service
- Model inference services
- Model training services
- Model repository services
- Agentic interfaces
Our first two pilots leveraging these features include:
- M3 Initiative AI Commons (with initial focus on cancer)
- Head and Neck Cancer Bank
Commons Services Operations Center (CSOC)
Our data commons and mesh platforms leverage the open-source Gen3 software. As part of Gen3, we are developing a new service called the Commons Services Operations Center (CSOC), which will enable third party organizations to more easily setup, configure, operate, and monitor their own Gen3 data commons over their own data and with their own LLMs. Over the next 3 years, CTDS is working to train approximately 25 groups and organizations to build and operate 3 or more Gen3 data commons each to create 50+ new Gen3 data commons to serve as a distributed platform for federated and distributed ML and AI.